aiMotive ships production RTL for neural processor
aiMotive in Hungary has shipped production-validated RTL of its fourth generation neural processing unit (NPU) to lead customers.
The aiWare4 design uses wavefront memory processing (WFRAM) to get five times the performance of the previous generation with half the silicon area in the neural network accelerator core. The company is working with Korean chip maker Nextchip as well as sensor maker onsemi and Sony on automotive designs.
Related articles
- Nextchip licenses aiWare4 for automotive AI processor
- European tech for Sony’s Vision-S electric car – video
- ON Semiconductor partners with AImotive on sensor fusion
“Our aiWare team has relentlessly refined our production validation processes to enable us to deliver customer configurations at record speed to full automotive quality for aiWare4,” says Márton Fehér, SVP hardware engineering at aiMotive. “Our wavefront processing makes full use of our new WFRAM technology, with many other architectural advances from aiWare3, to achieve exceptional PPA for our lead customers without compromising our leadership in high-efficiency execution up to 95% of the most demanding automotive inference CNN workloads”.
aiMotive was able to fine-tune the exact feature set of the aiWare4 production RTL to meet specific requirements from chip customers using a physical tile-based layout and dataflow methodology. This provides clock speeds for the production RTL of up to 1.3GHz over the full automotive AEC-Q100 Grade 2 temperature range on a 14nm process. The aiWare4 hardware IP has been assessed externally as suitable for certification to ASIL-B or higher as an SEooC (Safety Element out of Context, defined in the ISO 26262-10 standard).
The aiWare4 NPU scales from 1 to 256 TOPS and is supported by a comprehensive SDK featuring highly accurate offline performance estimation, enabling customers to accurately estimate and fine-tune their CNN workloads to within 5% of final silicon performance prior to first silicon.
aiWare4 hardware IP is available now for licensing.
Related articles
- Simulator supports multiple robotaxis
- AImotive, MathWorks team on AI simulation
- Self-driving car AI tech to power space electronics
- AImotive licenses neural network IP to NextChip
Other articles on eeNews Europe
- ARM launches highest performance Cortex M85 core in voice recognition subsystem
- Silicon Labs details hardware ML on wireless chip
- Tackling the challenges of RISC-V
- Espressif moves exclusively to RISC-V
- ARM launches its first Cortex-A pre-validated sub-system
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
